Bacteria Foraging Based Independent Component Analysis

Abstract

The present paper proposes a bacteria foraging optimization based independent component analysis (BFOICA) algorithm assuming a linear noise free model. It is observed that the proposed BFOICA algorithm overcomes the long standing permutation ambiguity and recovers the independent components(IC) in a fixed order which depends on the statistical characteristics of the signals to be estimated. The paper compares the performance of BFOICA algorithm with the constrained genetic algorithm based ICA (CGAICA) and most popular fast ICA algorithm. The proposed algorithm offers comparable or even better performance compared to fast ICA algorithm and faster convergence and better mean square error performance compared to CGAICA.